{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:3OM7TVIECNNAOWTHSVMPJ5LODV","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b4bc31e6a7cb08f7449de59dea7911ddf6ec5f714250dfa8fe0c62f0a339d008","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-13T21:47:08Z","title_canon_sha256":"01dbfa4f28b49de398720fbeb83ec222d9b30b1750fcb54ca639e850a7701962"},"schema_version":"1.0","source":{"id":"2607.20548","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.20548","created_at":"2026-07-24T00:23:22Z"},{"alias_kind":"arxiv_version","alias_value":"2607.20548v1","created_at":"2026-07-24T00:23:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.20548","created_at":"2026-07-24T00:23:22Z"},{"alias_kind":"pith_short_12","alias_value":"3OM7TVIECNNA","created_at":"2026-07-24T00:23:22Z"},{"alias_kind":"pith_short_16","alias_value":"3OM7TVIECNNAOWTH","created_at":"2026-07-24T00:23:22Z"},{"alias_kind":"pith_short_8","alias_value":"3OM7TVIE","created_at":"2026-07-24T00:23:22Z"}],"graph_snapshots":[{"event_id":"sha256:24cdbc0d56a8f38f9adc848cfd60002df9f4f8d10eae27a2d19b71f63f753723","target":"graph","created_at":"2026-07-24T00:23:22Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2607.20548/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Higher-order optimizers such as Muon and SOAP offer faster convergence than AdamW, but their computational cost and numerical stability challenges have limited adoption at scale. In this work, we adapt and enhance preconditioned gradient methods to overcome the practical challenges of large-scale LLM pretraining. We first identify instabilities in SOAP at large batch sizes and propose algorithmic modifications including per-step QR orthogonalization and improved preconditioning strategies that eliminate loss spikes and enable stable training in these regimes. We then present a unified empirica","authors_text":"Aditya Vavre, Boxiang Wang, Bryan Catanzaro, Deyu Fu, Dheevatsa Mudigere, Hao Wu, Jeff Pool, Michael Lightstone, Mikail Khona, Mike Chrzanowski, Mohammad Shoeybi, Mostofa Patwary, Nima Tajbakhsh, Tijmen Blankevoort","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-13T21:47:08Z","title":"SOAP, Muon, and Beyond: Pushing LLM Pretraining Scales"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.20548","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:acadceb32209521231a4d23a2070eb5a95a120c3a475741034d0ec2dd7434df7","target":"record","created_at":"2026-07-24T00:23:22Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b4bc31e6a7cb08f7449de59dea7911ddf6ec5f714250dfa8fe0c62f0a339d008","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-13T21:47:08Z","title_canon_sha256":"01dbfa4f28b49de398720fbeb83ec222d9b30b1750fcb54ca639e850a7701962"},"schema_version":"1.0","source":{"id":"2607.20548","kind":"arxiv","version":1}},"canonical_sha256":"db99f9d504135a075a679558f4f56e1d7b6bf61c5f5f8ded8da8afb14a7472bb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"db99f9d504135a075a679558f4f56e1d7b6bf61c5f5f8ded8da8afb14a7472bb","first_computed_at":"2026-07-24T00:23:22.055085Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-24T00:23:22.055085Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SfssFY8hnjI+S5o2VTYPJthsFGvx+cc52uOhFy5xvJ5F8goBk+RfsWv4BZDU+JOXy7WCtR1/glvhDXVK8uyUCA==","signature_status":"signed_v1","signed_at":"2026-07-24T00:23:22.056006Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.20548","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:acadceb32209521231a4d23a2070eb5a95a120c3a475741034d0ec2dd7434df7","sha256:24cdbc0d56a8f38f9adc848cfd60002df9f4f8d10eae27a2d19b71f63f753723"],"state_sha256":"3af0451ae472cde0715251dba6f867482db61eb3707f82f0db968ecbfc6c6b85"}